Japanese

2001

Application of Robust Gaussian Regression Filter
to Evaluation of Surface Roughness of Hardwood

The conventional filter of JIS and ISO for the evaluation of surface roughness has two disadvantages:

  1. End effect
  2. Ghost peaks are generated near deep valleys.
The Robust Gaussian Regression filter can solve these two problems.
In this study, we try to apply this filter for evaluating the surface roughness of hardwood.
Filtering Process of Robust Gaussian Regression Filter
the shape of the weight function Algorithm of the robust gaussian regression filter
Fig.1 Change of the shape of weight function

 Fig.2 Algorithm for iteration of the calculation of waviness profile
Waviness and Roughness Profiles obtained by Robust Gaussian Regression Filter(GR0)
waviness profile roughness profile
Fig.3 Waviness profiles obtained by JIS B 0632 and GR0
(Sanded surface of Japanese Oak:P240)
 Fig.4 Roughness profile obtained by JIS B 0632 and GR0
(Sanded surface of Japanese Oak:P240)
Relationship between grit numbers of coated abrasives and Ra or Rpk
Ra Rpk
Fig.5 Relationship between grit numbers of coated abrasives and Ra
(Sanded surface of Japanese Oak)
 Fig.6 Relationship between grit numbers of coated abrasives and Rpk
(Sanded surface of Japanese Oak)

Back to Projects
Home